Ilo E. Leppik MD
Director of Research, MINCEP Epilepsy Care, Minneapolis, Minnesota
Levetiracetam (LEV) is a new antiepileptic drug (AED) approved in Switzerland in late 1999 and in the United States in November 1999. Approvals in other countries have followed, and LEV is becoming more widely available. Most of the available literature on LEV is based on three studies performed for registration. Its efficacy has been tested mostly in localization-related epilepsies. However, it may have a broader applicability.
Early evidence of the clinical efficacy of LEV was obtained from several phase I and phase II studies (1, 2, 3, 4, 5, 6). These studies explored doses of 1,000 to 4,000 mg/day in various titration schedules. One study was an open dose-escalation model in 29 patients who received placebo for 4 weeks (baseline) followed by LEV 1,000 mg/day for 2 weeks, 2,000 mg/day for 2 weeks, 3,000 mg/day for 4 weeks, and, finally, 4,000 mg/day for 4 weeks. Twenty-seven of the 29 patients completed the study, and a substantially lower median seizure frequency was observed at all dosing periods. Placebo seizure frequency was 2.06 seizures per week, compared with 1, 1.5, 1, and 0.75 for the respective LEV doses. Somnolence and asthenia were most frequent with the 4,000-mg/day dose, suggesting that this may represent the upper limit in some patients (4). These early studies were influential in leading to the use of doses of 1,000 to 3,000 mg/day in the pivotal studies (7).
The efficacy of LEV as adjunctive therapy for localization-related epilepsy has been established by three pivotal, multicenter studies in the United States and Europe (8, 9, 10). These double-blind, placebo-controlled studies involved a total of 904 adult patients and constitute the core data for the approvals granted by various regulatory agencies. These studies also provided information regarding effectiveness for specific seizures types and quality-of-life issues. These three studies used similar patient populations and protocols, but differed in some details. Other studies on the effects of LEV are becoming available.
Pivotal Adjunctive Therapy Studies
One pivotal study (study 1) was performed in the United States (9) and two in Europe (8,10) (Table 42.1). The major differences between these involved dose regimens. In the U.S. study, final doses of LEV of 1,000 or 3,000 mg/day were compared with placebo. In the first European study (study 2), doses of 1,000 or 2,000 mg/day were compared with placebo (10). In the second European study (study 3), doses of 3,000 mg/day were compared with placebo, and titration to monotherapy was evaluated (8).
The U.S. study was a multicenter, randomized, add-on, double-blind, placebo-controlled, parallel-group trial that compared placebo, LEV 1,000 mg/day, and LEV 3,000 mg/day in patients with refractory partial epilepsy (9). Patients were enrolled from September 1994 to March 1996. The study consisted of a selection visit; a 12-week, single-blind, placebo baseline period; a 4-week doubleblind drug titration period; a 14-week double-blind treatment period; and an 8-week double-blind study medication withdrawal period, or the possibility of entering an open follow-up study. During the 12-week placebo baseline period, eligible patients stabilized on one or two appropriate AEDs at usual plasma concentrations were evaluated to determine whether they met all randomization criteria. Overall, 32% of patients were receiving one AED, 62% two AEDs, and 6% three or more AEDs. Patients had been stabilized on carbamazepine (57%), phenytoin (44%), gabapentin (28%), valproate (26%), phenobarbital (9%), primidone (6%), lamotrigine (5%), and “other” (7%) as one of their AEDs.
Randomization was in blocks by study site. Each patient was assigned a unique treatment number, which corresponded
to the blinded, randomized treatment, and each investigator was allocated treatment numbers in chronologic study entry order. Study medication was provided as identical white, film-coated, scored tablets containing either 166.5 mg LEV, 500 mg LEV, or placebo. LEV dosage was escalated at 2-week intervals during the 4-week, double-blind, dose titration period to the dose assigned at randomization. For the 1,000-mg/day group, dosages of LEV were 333 mg/day for 2 weeks, then 666 mg/day for 2 weeks and 1,000 mg/day started on the first visit of the evaluation period. The 3,000-mg/day group received 1,000 mg/day for two weeks, 2,000 mg/day for the next 2 weeks, then 3,000 mg/day. At the end of the 14-week treatment period, patients and investigators were given the option of entering a 1-year, open-label LEV continuation study. For patients choosing not to enter the follow-up study, LEV was withdrawn at a maximum of 500 mg/week during an 8-week study withdrawal period.
TABLE 42.1. OVERVIEW OF THE THREE PIVOTAL STUDIESa
The primary efficacy variable was the mean number of partial seizures per week computed over the entire 14-week evaluation period. Secondary efficacy variables were median percentage reduction compared with baseline, responder rate (number of patients with a minimum of 50% reduction from baseline in partial seizure frequency), and number of seizure-free patients.
Of 385 enrolled patients, 91 were determined to be ineligible and were never randomized. The most common reasons for not randomizing patients were failure to fulfill selection criteria (33 patients), consent withdrawn (19 patients), adverse events not related to study (14 patients), and protocol violation (12 patients). Of the 294 intent-to-treat (ITT; randomized) patients, 95 were in the placebo group, 98 in the LEV 1,000-mg/day group, and 101 in the LEV 3,000-mg/day group.
Of the 294 ITT patients, 285 completed the titration period, 268 (91.2%) completed the 14-week observation period, and 266 chose to enter the open-label, follow-up study. Premature study discontinuations during the dose titration and observation periods included 12 of 98 patients (12.2%) in the LEV 1,000-mg/day group, 8 of 101 patients (7.9%) in the LEV 3,000-mg/day group, and 6 of 95 patients (6.3%) in the placebo group.
The primary efficacy analysis of partial seizure frequency during the evaluation period was based on the 285 patients who completed the titration period. The percentage reduction in partial seizure frequency over placebo was 20.9% in the LEV 1,000-mg/day group and 27.7% in the LEV 3,000-mg/day group (p < .001 for both LEV groups).
Notably, reductions in partial seizure frequency over placebo were higher for both LEV groups at each evaluation visit (p ≤ .016). The median percentage reduction (and median absolute reduction) from baseline in partial seizure frequency also was greater in patients receiving LEV than placebo: 32.5% (0.81 seizures/week) with LEV 1,000 mg/day and 37.1% (0.98 seizures/week) with LEV 3,000 mg/day, versus 6.8% (0.13 seizures/week) with placebo (p < .001).
The 50% responder rate using the ITT population (294) was 37.1% (p < .001) for the LEV 1,000-mg/day group, 39.8% (p < .001) for the LEV 3,000-mg/day group, and 7.4% for the placebo group. The 75% and 90% responder rates calculated for the 285 patients who had completed titration are presented in Table 42.2.
Three patients receiving LEV 1,000 mg/day (not significant) and eight receiving 3,000 mg/day (p = .01) compared with none receiving placebo were seizure free during the entire 14-week evaluation period. The percentage of placebo-treated patients experiencing a >25% increase in partial seizure frequency was approximately twice that of LEV-treated patients: 13.8% (13 of 94 patients) for LEV 1,000 mg/day and 12.2% (12 of 98 patients) for LEV 3,000 mg/day, versus 25.8% (24 of 93 patients) for placebo (9).
Seizure reduction was notable for all seizure subtypes, with significant reduction in complex partial (type IB) seizures and secondarily generalized (type IC) seizures for the LEV 1,000- and 3,000-mg/day groups.
Of the 268 patients completing the study, 266 chose to continue in a long-term follow-up study. Of these, 148 were still on LEV 2 years later, and their median percentage
reduction in seizure frequency compared with their baseline was 73% at 2 years (9).
TABLE 42.2. RESPONDER RATES FOR THE THREE PIVOTAL STUDIES
This study was a randomized, double-blind, placebo-controlled, parallel-group study at 61 sites in Belgium, France, Germany, Luxembourg, Switzerland, and the United Kingdom (11). LEV was evaluated against placebo at doses of 1,000 mg/day (500 mg twice daily) and 2,000 mg/day (1,000 mg twice daily) as add-on therapy. LEV was started at 500 mg twice daily in both groups, and increased to 1,000 mg twice daily after 2 weeks in the 2,000-mg/day group.
The primary efficacy variable was the mean number of partial seizures per week computed during the evaluation period (i.e., seizure frequency). Secondary efficacy variables included the seizure frequency by seizure type and subtype, the responder rate, and the incidence of seizure-free patients.
Three hundred ninety-two patients were screened and 324 patients were randomized. All randomized patients became part of the ITT population: 112 in the placebo group, 106 in the group receiving LEV 1,000 mg/day, and 106 in the group receiving LEV 2,000 mg/day. Two hundred seventy-eight patients completed the study period. Dropout rates were 13% in the placebo group, 11% in the 1,000-mg/day group, and 18% in the 2,000-mg/day group. Across all treatment groups, the mean duration of epilepsy was 24 years and the mean age of epilepsy onset was 14 years. For more than half of the patients (57%), the cause of epilepsy was cryptogenic. At baseline, 31% of patients had simple partial seizures, 83% had complex partial seizures, and 26% had partial seizures with secondary generalization; 7% had generalized seizures or unclassified epileptic seizures. Baseline seizure frequency was comparable between groups. The number of AEDs taken by patients at baseline and throughout the study was similar among treatment groups. Most patients' conditions had been stabilized on carbamazepine (72%), phenytoin (22%), or valproate (21%). Among newer agents, the most frequently prescribed were vigabatrin (18%), lamotrigine (12%), and gabapentin (2%).
LEV significantly decreased the seizure frequency compared with placebo, with reductions over placebo of 16.4% for the 1,000-mg/day group [98% confidence interval (CI), 2.7% to 28.1%; p = .006] and 17.7% (98% CI, 4.1% to 29.4%; p = .003) for the 2,000-mg/day group. The median percentage reductions in seizure frequency from baseline were 6.1% for placebo, 17.7% for LEV 1,000 mg/day, and 26.5% for LEV 2,000 mg/day; no significant difference was identified between the two doses of LEV.
The median reductions in simple partial seizure frequency from baseline were 9.1%, 38.1%, and 46.3%, respectively, for the placebo group, the 1,000-mg group, and the 2,000-mg group. For patients with complex partial seizures, the median reductions from baseline in frequency of complex partial seizures were 9.1%, 12.4%, and 24.4%, respectively. Median secondarily generalized seizure frequency increased by 16.8% in the placebo group but decreased by 37.4% in the 1,000-mg group and 28.2% in the 2,000-mg group.
A significant difference in the responder rate was found between the placebo and treatment groups (p = .004). More patients treated with LEV 1,000 mg/day (22.8%, p = .019 versus placebo) or 2,000 mg/day (31.6%, p < .001 versus placebo) could be categorized as treatment responders than in the placebo group (10.4%). The number of patients needed to treat to get a responder with a ≥50% reduction in seizure frequency during treatment with LEV was 6.9 (95% CI, 4.3 to 17.9) for the 1,000-mg group and 3.5 (95% CI, 2.6 to 5.4) for the 2,000-mg group. In addition,
3.7% of patients in the placebo group experienced a ≥75% reduction in seizure frequency, compared with 10.9% (p = .03) for the 1,000-mg group and 16.8% (p = .001 for the 2,000-mg group. Five patients (5.0%) in the 1,000-mg group and two patients (2.0%) in the 2,000-mg group were seizure free during the evaluation period, compared with one patient (0.9%) in the placebo group who reported no seizures until study withdrawal at day 29.
This European, multicenter (47 institutions), randomized, double-blind, parallel-group, responder-selected study evaluated LEV 3,000 mg/day against placebo (8). Patients were recruited between June 1995 and May 1998, and a total of 343 were enrolled. This was a two-phase study, with evaluation of the target dose followed by withdrawal to monotherapy. Patients were evaluated on the target dose for 12 weeks (add-on evaluation period) after a 4-week titration period. During the following 2 weeks of the add-on evaluation period, investigators prepared a seizure profile of each patient using information obtained in the baseline and add-on phase and forwarded the profile to a central evaluator who determined which patients met the entry criteria for the monotherapy phase of the study.
The most frequently prescribed AEDs were carbamazepine (74%), lamotrigine (9%), valproate (8%), and phenytoin (6%). The main efficacy variable for the add-on phase of the study was seizure frequency, reported as the median number of partial seizures per week, and the responder rate (i.e., the proportion of patients with a reduction in partial seizure frequency of ≥50% compared with baseline).
Efficacy and safety analyses were conducted on the ITT population, which included all patients who were randomized and took at least one dose of study medication. For the monotherapy portion of the study, the primary efficacy assessment (i.e., the percentage of patients who completed the monotherapy phase relative to the number of patients randomized to study medication) was analyzed using Fisher's exact test.
Of the 343 enrolled, 286 patients were eligible for randomization to LEV or placebo (181 to LEV, 105 to placebo) and comprised the ITT study population. The percentage of patients completing the add-on phase was similar in both treatment groups (placebo, n = 90, 85%; LEV, n = 149, 82%). The median partial seizure frequency was significantly lower in patients treated with LEV (1.06 seizures/week) compared with placebo (1.75 seizures/week) during the add-on phase. Likewise, the median percentage reduction in partial seizure frequency from baseline to the add-on phase was 39.9% with LEV compared with 7.2% with placebo (p < .001). the responder rate was significantly higher with LEV than placebo (42.1% versus 16.7%, p < .001), with the odds of achieving a ≥50% reduction in seizure frequency 3.6 times greater. In addition, 8.2% (14/171) of LEV-treated patients remained seizure free during the add-on evaluation period, compared with only 1 patient in the placebo group (p = .012). The number of patients needed to treat to obtain one responder due to LEV effect was 3.9 (95% CI, 2.8 to 6.6) (12). The number of patients needed to treat to obtain one seizure-free patient due to LEV effect was 13.9 (95% CI, 8.5 to 37.4). Monotherapy results are discussed later in this chapter.
After the three pivotal studies were completed, a pooled efficacy analysis of the 589 LEV-treated patients (all doses and seizure types) and 310 placebo-treated patients who had baseline and add-on treatment data was done (13). Overall, there was a 31.3% median decrease in the number of seizures per week in all LEV-treated patients compared with a 5.4% decrease in the placebo group (p < .001). The overall ≥50% responder rate was 35.0% (206 of 589) for LEV-treated patients compared with 9.4% for placebo (p < .001). This responder rate was dose dependent, with the responder rate being 28.6% for the 1,000-mg/day dose, 35.2% for 2,000 mg/day, and 39.5% for 3,000 mg/day (13).
Additional evaluation was performed by seizure type: IA (simple partial), IB (complex partial), and IC (secondarily generalized), using the International League Against Epilepsy classification system (14). The efficacy of LEV on IC seizures over and above the ability to reduce all partial seizures was analyzed in all patients with IC seizures during baseline or the evaluation period (15). The ratio of secondarily generalized seizure frequency per week over partial seizure frequency per week was calculated for each patient both for the baseline and evaluation periods. Treatment was considered to have a greater ability to reduce IC seizures over and above all partial seizures if (Equation 1):
A logistic regression model taking into account treatment and study was used to compare treatment groups with respect to secondarily generalized seizure success rate. The median percentage seizure reduction from baseline was 42.7% for simple partial seizures, 36.1% for complex partial seizures, and 68.5% for secondarily generalized seizures (p < .05 for all seizure types versus placebo).
The percentage of patients successfully treated for secondarily generalized seizures (ratio less during evaluation versus baseline) was significantly greater in the LEV group (59%) than in the placebo group (44.9%). The odds of successful treatment of secondarily generalized seizures with LEV therapy were 1.83 (95% CI, 1.10 to 3.05) times higher than the odds of successful treatment with placebo (15).
The major study of monotherapy completed to date was the second phase of the second European study (study 3 in this chapter) (8). Only patients who responded to LEV or placebo during the add-on phase were eligible for entry into the monotherapy phase. The following criteria (relative to baseline) must have been met: (a) ≥50% reduction in partial seizures or ≥35% reduction in simple partial seizures (type IA) provided that complex partial seizures (type IB) were reduced by ≥50% and secondary generalized seizures (type IC) were not higher than baseline; (b) no doubling of IB or IC seizure frequency; and (c) no type IC seizures during add-on phase if not present during baseline.
In the first half of the monotherapy phase, the dose of the standard AED was withdrawn gradually over a period of up to 12 weeks (downtitration period), while the dose of the study medication remained constant. The latter half was the monotherapy evaluation phase, in which patients received only study medication for up to 12 weeks. Patients were withdrawn from the monotherapy phase if they met one of the following escape criteria relative to baseline: (a) doubling of monthly type IB or IC seizure frequency, (b) occurrence of status epilepticus, or (c) emergence of type IC seizures if none occurred during baseline.
For ethical reasons, the study was designed under the assumption that no more than 10% of patients taking placebo would fulfill the responder selection criteria. When a total of nine placebo responders was reached, for ethical reasons, each subsequent placebo responder was switched to LEV for the monotherapy phase without breaking the blind. These patients were analyzed as if they were still taking placebo.
The primary efficacy variable of the study was the percentage of patients who completed the monotherapy phase relative to the number of patients randomized to receive study medication. Except for the primary efficacy assessment, no statistical comparison between treatment groups was performed in the monotherapy phase because the placebo group was established solely to maintain the double-blind status of the study and not to be used as a comparative group.
Of the 239 patients who completed the add-on phase, 86 patients were eligible to enter the monotherapy phase of the study (placebo, n = 17, 16.2%; LEV, n = 69, 38.1%). For ethical purposes, 8 of the 17 patients taking placebo were switched to treatment with LEV, but for analysis, they remained in the placebo group. Twenty-five patients (placebo, n = 5; LEV, n = 20) were withdrawn during the downtitration period mostly because they met escape criteria (placebo, n = 2; LEV, n = 11). Forty-nine of the 69 (70%) patients receiving LEV were successfully downtitrated to LEV monotherapy. In these patients, the median absolute reduction in partial seizure frequency from baseline to the monotherapy evaluation period was 0.61 (p = .012), with a 73.8% median percentage reduction in partial seizure frequency (p = .037).
Of the ITT population, 19.9% (36/181) of patients randomized to LEV treatment completed the study, compared with only 9.5% (10/105) in the placebo group (p = .029). The odds of completing the study were 2.36 times higher for LEV than placebo. Of note, 4 of the 10 patients in the placebo group who completed the study had been switched to treatment with LEV at the start of the monotherapy phase for ethical reasons.
Compared with the add-on phase, a slightly higher frequency of seizures was reported during the monotherapy evaluation period (median increase = 0.04). In addition, the 50% responder rate during this treatment period was 59.2%. Nine of 49 patients (18.4%) taking LEV were completely seizure free throughout the 12-week monotherapy evaluation period. Of these nine patients, six had been seizure free since the uptitration period, a total interval of 31 to 39 weeks. In the placebo group, three patients were seizure free throughout the 12-week monotherapy evaluation period. One of these patients had received placebo during the add-on phase and became seizure free during the 12-week monotherapy evaluation period subsequent to his switch to LEV in the monotherapy phase.
Other Seizure Types
Animal studies suggest that LEV should have a broad spectrum of action in human epilepsy (16). However, few clinical reports are available yet. One review of participants of various open-label studies of LEV found 13 who had juvenile myoclonic epilepsy and had not responded to valproate and lamotrigine. Twelve had suppression of seizures by 50% or more, and 6 of 13 became seizure free for the duration of the various studies, with a mean of 36 months (17). In one study of photosensitive epilepsy, LEV was found to be effective in blocking the induced epileptiform response (5). Because LEV is structurally related to piracetam, it has been suggested that it may be effective in myoclonic syndromes (18). To date, six patients with myoclonus treated with LEV have been reported. In one report, three patients with debilitating myoclonus (one with postanoxic myoclonus, two with Unverricht-Lundborg disease) were treated with LEV 4,000 mg/day as the initial dose with no titration. All three patients had a dramatic reduction of myoclonus. However, because LEV was an investigational drug, it was replaced by piracetam (19). Three additional patients with posthypoxic or postencephalitic myoclonus successfully treated with LEV have been reported (20).
CLINICAL USE ISSUES
Initiation of Treatment
In most studies, LEV had been titrated to higher dosages. In the U.S. phase III studies, the treatment groups were divided into low-dose titration (333 mg/day, 666 mg/day,
and 3,000 mg/day) or high-dose titration (1,000 mg/day, 2,000 mg/day, and 3,000 mg/day). There appears to be no differences in dropout rate or adverse event reports between these groups. One study was specifically designed to address this issue and involved 119 subjects who received 1,000 mg twice daily (2,000 mg/day) or 2,000 mg twice daily (4,000 mg/day) immediately. The overall responder rate (≥50% reduction in seizures compared with baseline) was 43.0% (21). Both 2,000-mg and 4,000-mg doses were well tolerated, with similar side effect profiles and discontinuation rates. Labeling in the United States recommends initiation with 500 mg twice daily with dosage increases based on clinical response (22). However, in clinical practice involving less refractory patients, some individuals do not tolerate these initial dosages and slower titration may be needed (Leppik, personal observation).
Onset of antiseizure activity is rapid after the initiation of LEV. Data from the three phase III trials were evaluated to determine the proportion of seizure-free days after initiation of treatment. The percentage of patients having a seizure-free day was increased by 11% to 20% compared with placebo (p < .001) 1 day after initiation, and similar increases were seen if a 3- or 7-day period was used (23). A similar rapid onset was observed for the patients with myoclonus (19).
Studies performed for the purpose of registration have narrow entry criteria that often exclude or limit children, women of childbearing potential, and persons with medical illnesses. Some information regarding these populations has been obtained from various smaller open-label studies (24).
A safety, tolerability, and pharmacokinetic study of LEV in pediatric patients with epilepsy was conducted at six sites in the United States (25,26). Twenty-four patients were enrolled (15 boys, 9 girls) with a mean age of 9.4 years (range, 5.6 to 12.6 years), and all had partial seizures with or without secondary generalized seizures. In this add-on study, all patients were taking a steady regimen of one other AED. LEV was administered as a single dose of approximately 20 mg/kg on day 1 of active treatment and then, starting on day 2, LEV was given every 12 hours beginning at approximately 10 mg/kg/day. Dosages were escalated to ~40 mg/kg/day (in equal doses twice daily), in 2-week increments, with a safety evaluation before each dosage escalation (25). The 40-mg/kg/day dosage would equal 3,000 mg/day LEV in an adult patient weighing approximately 75 kg.
Pharmacokinetic data were obtained after the initial administration of 20 mg/kg (26). Titration was followed by an 8-week evaluation period. The clearance of LEV was 30% to 40% higher in these children than in adults, and the half-life was correspondingly shorter (~6 hours). The preliminary recommendation from this study is that the maintenance dose of LEV in children should be 30% to 40% higher on a weight basis than the recommended maintenance dose in adults. Although efficacy was observed, results from ongoing studies involving larger numbers of subjects of the efficacy of LEV in children are needed.
One study was conducted in 16 hospitalized elderly volunteers (5 men and 11 women) (11). Plasma concentration and urinary excretion of LEV were studied after a single 500-mg dose and also after 500 mg twice daily for 10 days. Creatinine clearance (ClCr) was diminished but higher than 30 mL/min/1.73 m2 in all subjects (range = 30.1 to 73.9 mL/min/1.73 m2). After a single 500-mg oral dose of LEV, plasma half-life was longer (10.3 ± 1.7 hours) compared with that seen in young adult volunteers (7.7 hours). After multiple dosing, plasma half-life was 10.4 ± 1.8 hours.
Results indicated that although clearance was reduced and half-life prolonged in these elderly subjects, these effects were entirely attributable to reduced renal function.
Three studies have been done to assess LEV in patients with impaired renal function: a single-dose study, a multiple-dose study, and a study in patients with end-stage renal disease. A single-dose study of LEV was conducted in 11 subjects (6 men and 5 women) with varying degrees of renal function (11). Two of the subjects had normal renal function (CICr ≥ 90 mL/min/1.73 m2). The other nine subjects had mild, moderate, or severe impairment (CICr range = 5.6 to 84 mL/min/1.73 m2). The dose of LEV was 500 mg. The renal clearance of LEV was directly proportional to the CICr. The half-life of the drug ranged from 10.4 hours in those with mild impairment to 24.1 hours in those with severe impairment.
In a follow-up to the single-dose study, patients with renal insufficiency were administered multiple doses of LEV (11). Twenty-one patients were enrolled (14 men and 7 women), 5 of whom had normal renal function (CICr ≥ 90 mL/min/1.73 m2). The remaining patients had varying degrees of renal impairment (ClCr = 13 to 80 mL/min/1.73 m2). The mean age was 57.9 years (range, 21 to 77 years). A single dose of LEV was administered on the first day of the study. Patients with ClCr ≥ 40 mL/min/1.73 m2 received a single 1,000-mg dose of LEV; subjects with poorer renal function received a single dose of 500 mg. On days 2 to 9, patients with ClCr ≥ 40 mL/min/1.73 m2 received a 2,000-mg dose of LEV (1,000 mg twice daily), whereas subjects with poorer renal function received 1,000 mg (500 mg twice daily). Total-body clearance was reduced
in patients with renal impairment and was proportional to ClCr. Multiple dosing did not affect the pharmacokinetic parameters.
A study of five patients with end-stage renal disease undergoing dialysis (two male and three female; mean age = 58.4 years, range = 36 to 72 years) also was conducted (11). A single dose of 500 mg of LEV was administered to each subject immediately after a 4-hour dialysis. As expected, the half-life of the drug was prolonged, to an average of 24.6 hours, and the principal inactive metabolite of LEV, UCB L057, accumulated during the interdialysis period. After dialysis, the plasma LEV and metabolite concentration was reduced by approximately 50%.
These three studies in patients with impaired renal function indicate that LEV dosing must be individualized according to the patient's renal function status.
Follow-up evaluations of persons participating in various LEV studies have provided information regarding longterm efficacy. A total of 1,442 patients with refractory epilepsy (median age 36.0 years, range 7 to 78 years) were exposed to LEV. The mean duration of treatment was 537 (standard deviation, 492) days, and the median LEV dosage was 3,000 mg/day. The LEV retention rate was estimated to be 58% after 1 year, 43% after 2 years, and 36% after 3 years. In the 970 patients with a baseline evaluation, the median seizure reduction under LEV treatment was 37.8% (evaluation period, 1 week to 5 years). At least a 50% seizure reduction was seen in 38.5% of patients, and 19.2% had at least a 75% seizure reduction. From the 599 patients with an exposure to treatment exceeding 1 year, 67 (11.2%) because seizure free for at least 1 year. No tolerance to the effects of LEV has been noted (23).
Data of all patients with epilepsy exposed to LEV during the developmental program of the drug were analyzed. The retention rate was analyzed using the Kaplan-Meier method. A Cox regression model was used to identify predictors of a higher retention. Data collected through a cutoff date of June 30, 1999 have been presented (27). A total of 1,422 patients with refractory epilepsy were exposed to LEV (median age, 36 years, range 5 to 78 years); the median dose was 3,000 mg/day. The mean duration of exposure was 622 days (range, 1 to 2,984 days). Forty percent of patients were still being treated at the end of the observation period. In 16% of patients, the treatment was terminated because of adverse events, in 18% because of lack or loss of efficacy. The LEV retention rate was estimated to be 60% after the first year, 37% after 3 years, and 32% after 5 years. The median seizure reduction from baseline was 40%. Thirty-nine percent of patients had a seizure reduction of 50% or more, 20% of patients of 75% or more. Thirteen percent of patients were seizure free for at least 6 months and 8% for at least 1 year in ITT analysis. If the analysis was restricted to patients who effectively were exposed to LEV treatment exceeding 6 months, seizure freedom for at least 6 months was observed in 17% of patients. The drug was usually well tolerated, and adverse events were mainly central nervous system related and mild, including somnolence and dizziness. There was no evidence for idiosyncratic side effects. No tolerance to the effects of LEV has been noted. The Cox regression model identified four factors that were correlated with higher retention rates: (a) higher maintenance doses of LEV, (b) lower starting doses of LEV, (c) presence of convulsive seizures, and (d) lower number of concomitant AEDs. Factors that did not significantly influence retention on LEV were age at onset of epilepsy, duration of epilepsy, history of status epilepticus, and history of withdrawal seizures. These data indicate that LEV is an effective and well tolerated new AED. The long-term retention rate and percentage of seizure freedom in patients with refractory epilepsy seem to be higher than those of other new AEDs.
SURROGATE MEASURES OF EFFICACY
Quality of Life
Although efficacy as measured by seizure frequency is the principal outcome evaluated in studies, of great importance to the patient is the change in quality of life made possible by a new treatment. The ultimate effectiveness of a drug when used in typical clinical situations encompasses the patient's perception of its modulation of the overall health status. The U.S. phase III study incorporated the QOLIE-31 battery (31-item Quality of Life in Epilepsy) and was completed by 246 of the 294 patients enrolled (28). Overall, “seizure-worry” cognitive functioning, and overall QOL scores improved compared with placebo (p < .0003; p < 0.04; and p < 0.01, respectively). The improvements were most marked for persons who were responders (28). A shorter form, QOLIE-10, compares well with the longer version and may be more useful in clinical settings (12).
Another measure of improvement is the evaluation of seizure-free days. Of 904 patients, 846 with evaluable data were considered in the analysis of seizure-free days: 189 at 1,000 mg/day, 90 at 2,000 mg/day; 269 at 3,000 mg/day, and 298 on placebo. A seizure-free day was defined as a reported day without seizures. The mean proportion of seizure-free days per year was defined as the proportion of days without seizure computed over a period of time and extrapolated to a per year value. The increase in days without seizure was examined during the evaluation period and compared with baseline. Patient diaries were used as data sources for seizure dates and characteristics. Treatment
effect was tested in each individual study and in a pooled analysis, using an analysis of variance with baseline as covariate. Results were confirmed by a nonparametric Wilcoxon test. In the three individual studies and in the pooled analysis, the increase in seizure-free days was statistically significant at all LEV doses over placebo. Considering the impact on days with seizure, the pooled analysis reveals that from an adjusted baseline of 105.8 seizure days per year, patients experienced a 14.7-days/year (13.9%) decrease for LEV 1,000 mg/day, a 19-days/year (17.9%) decrease for LEV 2,000 mg/day, and a 20.2-days/year (19.1%) decrease for LEV 3,000 mg/day in the number of days with seizures (p < .0001 for analysis of covariance). The treatment effect was confirmed by the results of the Wilcoxon test (p < .0001) (29).
Levetiracetam has proven efficacy in the treatment of localization-related epilepsies. Dosages of 1,000 to 3,000 mg/day have been thoroughly evaluated in placebo-controlled trials. In clinical use, higher dosages may be found to be useful. At present, LEV does not have a monotherapy indication in the United States, but in some persons, withdrawal to monotherapy may be a useful approach. It appears to maintain antiseizure activity for extended periods. Further studies in absence and myoclonic disorders are needed, but preliminary observations are favorable for broader use (30).
Supported in part by NIH grant P50 NS16308. The author thanks Liliane Dargis for word processing.